Miaohua Zhang

Miaohua Zhang is a computer vision researcher at CSIRO in Canberra whose work covers robust image representation, few-shot learning and self-supervised models for plant disease detection. Her studies ran from a B.Eng. in 2011 and an M.Eng. in 2014 at Xi'an University of Technology to a PhD at Griffith University in 2019. In 2020, she introduced Corr-Tensor, a robust tensor decomposition method coauthored with Yongsheng Gao, Changming Sun and Michael Blumenstein, while working as a research fellow in Griffith's ARC Research Hub for Driving Farming Productivity and Disease Prevention from May 2020. She joined CSIRO's Efficient Computer Vision team in January 2023, coauthored SSFE-Net on few-shot class incremental learning for WACV 2023, and worked on the Vision Mamba models ConMamba and PSMamba in 2025. She is an early career research fellow in that team.

Now at: CSIRO

Timeline

2011–2019

Degrees from Xi'an to Griffith

She earns a B.Eng. in computer science and technology from Xi'an University of Technology in 2011 and an M.Eng. in pattern recognition and machine intelligence there in 2014. She completes a PhD at Griffith University in Queensland, Australia, in 2019.
2013–2014

Two IEEE image-processing papers

She coauthors "Robust ellipse fitting based on sparse combination of data points" in IEEE Transactions on Image Processing in 2013. In 2014, "Distributed dictionary learning for sparse representation in sensor networks" appears in the same journal.
May 2020

Fellow at a Griffith research hub

In May 2020, she becomes a research fellow in the ARC Research Hub for Driving Farming Productivity and Disease Prevention at Griffith University.
2020

The Corr-Tensor method

In 2020, she coauthors "Robust Tensor Decomposition for Image Representation Based on Generalized Correntropy" with Yongsheng Gao, Changming Sun and Michael Blumenstein, published in IEEE Transactions on Image Processing. The paper proposes Corr-Tensor, a tensor decomposition method built on a generalized correntropy criterion that stays robust to outliers.
January 2023

Joining CSIRO in Canberra

In January 2023, she joins CSIRO's Efficient Computer Vision team at the Black Mountain site in Canberra as an early career research fellow. Her research covers domain adaptation, self-supervised learning and few-shot learning, applied to object detection, image segmentation and image classification.
2023

SSFE-Net at WACV

In 2023, she coauthors SSFE-Net with Zicheng Pan, Xiaohan Yu and Yongsheng Gao at the IEEE/CVF Winter Conference on Applications of Computer Vision. The method combines self-supervised learning and knowledge distillation to strengthen a low-capacity backbone for ultra-fine-grained few-shot class incremental learning, and it comes with new benchmarks for that task.
2024

A plant disease review

In 2024, she coauthors "Plant disease detection using self-supervised learning: A systematic review" in IEEE Access. She also coauthors "Pseudo-set frequency refinement architecture for fine-grained few-shot class-incremental learning" in Pattern Recognition.
As of October 2026

Vision Mamba for plant disease

In 2025, she coauthors ConMamba, a self-supervised Vision Mamba framework for plant disease detection, and PSMamba, a progressive self-supervised framework for plant disease recognition. ConMamba appears in Pattern Recognition in 2026. She works at CSIRO in Canberra.